are virtual reference services color blind?

Library & Information Science Research 28 (2006) 501–520

Are virtual reference services color blind?

Pnina Shachaf ⁎, Sarah Horowitz

School of Library and Information Science, Indiana University, 1320 East 10th Street, LI005A, Bloomington,IN 47405-3907, USA

Available online 1 November 2006

Abstract

This study reports an experiment that examines whether librarians provide equitable virtualreference services to diverse user groups. The relative absence of social cues in the virtual environmentmay mean greater equality of services though at the same time greater inequalities may arise aslibrarians can become less self-aware online. Findings indicate that the quality of service librariansprovide to African Americans and Arabs is lower than the quality of service they provide to Caucasian,Hispanic, Asian, and Jewish students. This study adds to the knowledge of subjective bias in thevirtual environment by specifying those that are discriminated against online, identifying the kinds ofdiscriminatory actions of virtual reference librarians, and identifying the type of queries that morefrequently result in unbiased service.© 2006 Elsevier Inc. All rights reserved.

1. Introduction

Virtual reference, which allows users to connect easily with librarians online, is becominghighly popular. Librarians answer thousands of questions every day over the Internet (Coffman& Arret, 2004). Although the service is maturing much more research is still needed. Inparticular, there is a need to gain a better understanding of the role of virtual reference inproviding services to diverse and cross-cultural user groups (Shachaf, Hara, Meho, Kwon, Li,

⁎ Corresponding author. Fax: +1 812 855 6166.E-mail addresses: [email protected] (P. Shachaf), [email protected] (S. Horowitz).

0740-8188/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.lisr.2006.08.009

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.lisr.2006.08.009

502 P. Shachaf, S. Horowitz / Library & Information Science Research 28 (2006) 501–520

Sakai, & Lankes, 2005). This research gap is now becoming more important with the increasednumbers of diverse and non-traditional students attending colleges and universities.

Today more than ever minorities and international students attend higher educationinstitutions in North America (Curry & Copeman, 2005; Whitmire, 1999, 2003). These diverseuser groups are making growing use of library and reference services (Curry & Copeman,2005; Liu & Redfern, 1997; Whitmire, 1999, 2003; Zoe & DiMartino, 2000). They use theacademic library heavily (African American students use the library more thanWhite students)and request more assistance in searching databases (non-native English speakers comparedwith native English speakers) (Whitmire, 1999, 2003; Zoe & DiMartino, 2000). They may alsoheavily use virtual reference services.

The provision of online reference services, specifically via e-mail, enables users who wereuncomfortable approaching a librarian in person to submit anonymous questions to a virtualreference service (Coffman, 2003). Likewise, certain user groups do not come to the librarydue to physical disabilities, scheduling constraints, or geographical distance. These users toomight find virtual reference to be more accessible for their needs, as it provides services tousers any time, any place. Users can send reference queries via e-mail whenever they havequestions (Bushallow-Wilbur, Devinney, & Whitcomb, 1996; Stacy-Bates, 2003).

However, providing reference services to diverse user groups is a challenging task thatinvolves overcoming many barriers (Curry & Copeman, 2005). The American LibraryAssociation's (1995) code of ethics guides reference librarians to provide unbiased responsesto all requests and users. Additionally, Section 201(a) of the Civil Rights Act of 1964 (42 USC§ 2000a(a) (1964)) specifies that “all persons shall be entitled to the full and equal enjoymentof the goods, services, facilities, privileges, advantages, and accommodations of any place ofpublic accommodation, as defined in this section, without discrimination or segregation on theground of race, color, religion, or national origin.” Yet it is possible that reference librarians inacademic libraries provide biased reference service. Providing services to diverse user groupsover the Internet (virtual reference) may increase or decrease such bias.

It is possible that reference services will be more equal when provided online becauseduring the reference interview (via e-mail) blatant differences do not interfere. Also, e-mailmediates intercultural communication and reduces miscommunication (Shachaf, 2005). Whencommunication is mediated by computers, social presence is reduced (Sproull & Keisler,1986). Interactions are more impersonal and task oriented because of the lack of social cues,lack of context, and lack of non-verbal behavior. The decrease in social cues can have ademocratizing effect on communication resulting, for example, in status equalization. Therelative absence of social cues may mean greater equality of services in the virtualenvironment. Due to the lack of social cues in e-mail-mediated communication, virtualreference can facilitate the provision of unbiased service to diverse user groups.

E-mail bears many challenges for librarians in the provision of reference services, however.Primarily due to the lack of non-verbal cues during the reference interview, misunderstandingcan occur (Abels, 1996). Furthermore, discrimination is more likely to be expressed overtlydue to the anonymous, spontaneous, impersonal, and uninhibited nature of computer-mediatedcommunication (Glaser & Kahn, 2005); librarians can express themselves in less self-conscious and socially desirable ways. Although librarians will not be likely to deny some

503P. Shachaf, S. Horowitz / Library & Information Science Research 28 (2006) 501–520

resources or services on the basis of group membership, they may find an excuse to behavediscriminatorily at the moment. In computer networks, for example, higher levels of hostile,uninhibited, and flaming behavior have been observed (Douglas & McGarty, 2001). Douglasand McGarty (2001) claimed that it is likely that in the virtual environment subjective bias willbe similar to the pre-civil rights era and that greater inequality will arise. This discriminationcan be seen as similar to that found in studies from the pre-civil rights act era that found formalwritten requests for service to be rejected more than in person requests by minorities (Glaser &Kahn, 2005). In providing virtual reference services, librarians can become less self-aware andless likely to monitor their behavior and therefore more likely to react on impulses that wouldnormally be inhibited. Thus, it is also possible that greater inequalities will arise. While it ispossible that inequality in reference services will increase in the virtual environment due to thederegulating effect it has, it is likewise possible that librarians may be able to provide unbiasedservice online. This study addresses the lacuna in studies that examine discrimination online(Glaser & Kahn, 2005). The study aims to add to the knowledge of subjective bias in thevirtual environment. Specifically, it addresses the need for more studies that examine servicediscrimination by librarians and aims to identify the kinds of discriminatory actions of virtualreference librarians. This study focuses on the following research question: Are virtualreference services color and gender blind and are they providing unbiased services to diverseuser groups? More specifically, this study clarifies who is more likely to be discriminatedagainst online by virtual reference librarians.

2. Procedures

An experiment using scenarios of information needs was conducted. The experimentalfeature of the study is that the requests have six versions that differed only in the implicitethnicity of the user involved in the six incidents. Each version of the request represented oneethnic group: African American, Hispanic, Asian, or Caucasian; or one religious identification:Christian (Caucasian), Muslim, or Jewish. The targeted reference service received a version ofthe same type of request but with a different user name (indicating a different ethnic group orreligious affiliation). In this way, it was possible to determine whether librarians provideequivalent service to different groups when salience of diversity (e.g., ethnic group) is not anobvious factor and when all other factors are constant.

This method is in particular relevant when attitude and behavior variability are measured onsensitive variables, such as diversity. If people avoid blatant discrimination to avoid socialramifications they might monitor their discriminatory behavior closely if they believe the studyis related to race (Bushman & Bonacci, 2004).

There are many ways that have been applied to evaluate reference services; these havefocused on the types of questions asked at the desk, the accuracy of the information providedby the reference librarian in her reply, user satisfaction with the reference service, librarian'sbehavior, and library collections (Gross & Saxton, 2002). Some researchers applied the samemethods to evaluate virtual reference services while others advised the development and use ofnew ways to evaluate virtual reference services (Hernon & Calvert, 2005; McClure, Lankes,


Gross, & Choltco-Devlin, 2002; White, 2001). At the same time, the need to conduct moreevaluative studies of virtual reference service has likewise been emphasized (White, 2001).

Unobtrusive methods for traditional reference services evaluation were used for decades.One of the most well known unobtrusive studies was conducted about two decades ago andgenerated many replications which consistently found that librarians' responses are accurateonly 55% of time (Hernon & McClure, 1986). Whitlatch (1989; 2001) emphasized that thismethod can be used effectively to evaluate virtual reference as well.

This study applied an unobtrusive method, where responses to queries are assessed withoutthe reference service providers knowing that they are being studied. Several studies reportfindings from unobtrusive studies that assess the quality of virtual reference services (Carter &Janes, 2003; Kaske &Arnold, 2002; Stacy-Bates, 2003). The assumption is that, if an obtrusiveapproach were employed, librarians would more likely try to act in a fair and impartial waybecause the potential for discrimination and inequitable service would be obvious.

3. Data collection

During summer 2005, all Association of Research Libraries (ARL) member libraries wereinvited by an e-mail sent to the head of the reference department to participate in the study.Two batches of requests were sent and 23 libraries agreed to take part in this study—for aparticipation percentage of 19%. These and other libraries raised a concern that their servicesto unaffiliated users (users who are not affiliated with the institution) are limited. “Acceptingquestions from non affiliates of the library has both pros, such as good public image, andcons, such as diverting reference service providers' time from serving affiliates” (Stacy-Bates,2003, pp. 60–61). Another frequent concern that librarians have raised when asked toparticipate was that they cannot decide to participate due to the fact that they provide virtualreference as part of a collaborative effort. Despite the fact that the recruitment letter and theinformed consent form specified that the libraries which agreed to participate would receive e-mail queries, many libraries interpreted virtual reference service as primarily chat or real timeservice, and not e-mail or Web form. However, White (2001) defines virtual reference serviceas “an information access service in which people ask questions via electronic means (such ase-mail and Web forms). In turn knowledgeable individuals answer the questions, andresponses are transmitted via electronic means” (p. 211). Following Stacy-Bates' (2003)definition, virtual reference service in this study refers to e-mail: either a mail-to link or a Webform that users can fill out to ask reference questions.

In the fall of 2005, e-mail reference requests were sent to the 23 libraries; each libraryreceived one request per week during six consecutive weeks. Each library received fivedifferent types of requests using six different names and six different e-mail accounts. Everyweek, 23 messages were sent from one e-mail account (one user), one per library, five fromeach type of question. The chronological order of the users' messages that was used wasdefined by alphabetic order of user's last name. Using this order, it was possible to overcomethe chance that current events might confound our results. For example, just before the first 23messages were about to be sent, during the second week of September 2005, racial and class


issues were discussed in the media following Hurricane Katrina's devastation of New Orleans.During the same week, Israel withdrew from Gaza. These two events could potentiallyinfluence the results of the study. Any informed decision about the order of the users couldpotentially threaten the reliability of the results. For that reason, a randomly arranged order(alphabetical order of user's last name) was used for data collection.

The names that were used in this study were selected from lists of names (or baby names)that are available online. The names that represent ethnic groups are Latoya Johnson forAfrican-American, Rosa Manuz for Hispanic, Chang Su for Asian (Chinese), and MaryAnderson for Caucasian. Representing religious group affiliation, the names that were used areMary Anderson for Christian, Ahmed Ibrahim for Muslim, and Moshe Cohen for Jewish. Atotal of six names were included; four of them are female names and two are male names.

Five reference queries were used in this study; three of them followed Stacy-Bates' (2003,p. 61) examples, which according to her study are expected to be answered by more than 90%of the libraries. Following these findings, it was expected that most of the participant ARLlibraries will respond to unaffiliated users on these three queries. However, it was alsoexpected that academic libraries will not be inclined to answer queries 4–5 when made byunaffiliated users. When the virtual reference service policies of the participant libraries wereexamined, it was found that most of them limit their virtual reference services to unaffiliatedusers. These policies specify that the library will answer unaffiliated users' requests only whenthey submit questions that are specific to the library or university. Assuming that referencelibrarians are aware of the policies for unaffiliated users, it was expected that they will notprovide support to unaffiliated users with their topical requests. Finally, it was assumed thatreference librarians will comply with intellectual property and copyrights laws and will followthe terms of use of electronic journals and databases. Therefore, it was expected that the fifthrequest will be rejected by the participant libraries. These are the queries that were sent:

1. Dissertation query (Stacy-Bates, 2003): Can you tell me the title of [name]'s dissertation?[She/He] finished [his/her] degree at [institution name] in [year]. Do you have it in yourlibrary? How can I obtain a copy of this dissertation?

2. Sports team query (Stacy-Bates, 2003): How did [sports team name] become the name for[institution name]'s sports teams? Can you refer me to a book or article that discusses it?

3. Population query (Stacy-Bates, 2003): Could you tell me the population of [institution'scity name] in 1963 and 1993?

4. Subject query: Could you help me find information about [special collection topic]? Canyou send me copies of articles on this topic?

5. Article query: Can you send me by e-mail a copy of the article “Free Indirect Discourse andNarrative Authority in Emma” by D.P. Gunn?

Before the questions were sent, information about each institution was collected into a file.This information was later used when preparing the specific questions about informationunique to the institution. Each institution file included (1) details of a dissertation that wasawarded by the institution in 1964 (or the earliest date thereafter) which was identified usingDissertation Abstracts, (2) sports team's name as identified from a search on the parent


institution Web site, (3) topics of special collections for the subject query, and (4) policiesabout service restrictions for unaffiliated users and response time. In addition, a full-text articlethat was available via EBSCO Academic Search Premier was selected for the fifth question. Itwas assumed that any of the ARL libraries will have access to the full-text of this article, butthat due to licensing restriction librarians will be less likely to send the full text of the article.

A total of 138 queries were sent during September and October 2005. The counterbalancedmethod was used to avoid variables confounding. The technique in counterbalancing is tomake sure that each user name appears in each position an equal number of times. Each usersent the same number of messages, each library received only one request from each user, andeach library received a specific type of request only once. Each week, a different question wasreceived at the reference service of an institution from a different user.1

Each of the repeated requests, which were sent during the sixth week, was different from thefirst request an institution received during the first week of data collection. These requests had adifferent article title, dissertation title, and years but followed exactly the same text.

Because it was clear that if the requests were to be sent to the library on the same day of theweek and at the same hour of the day librarians would be suspicious about this pattern of requestscoming in from unaffiliated users, the requests were sent during different days and hours. As aresult, some of the messages were sent during the weekend. This created some difficulties in theinterpretation of the findings later, such as interpreting the results for response time.

4. Data analysis

All 138 queries were uploaded into Nvivo2.0. Nvivo is a QSR software that supportsqualitative analysis. Using Nvivo facilitates content analysis and the search for frequencies andco-occurrences of codes and attributes. Because the transactions' content categories had to becorrelated with the attributes of each transaction, such as user name, response time, length ofreply, and type of question, to name a few, the use of Nvivo was instrumental. Further, Nvivomatrix capabilities facilitate the identification of patterns among categories and also betweencategories and attributes.

1 An example of the chronological order, user name, and type of questions a particular institution received as partof this study is given below.Week 1: Mary Anderson—Could you tell me the population of [city name] in 1963 and 1993?Week 2: Moshe Cohen—Could you help me find information about [special collection topic]? Can you send mecopies of articles on this topic?Week 3: Ahmed Ibrahim—Can you send me by e-mail a copy of the article “Free Indirect Discourse and NarrativeAuthority in Emma” by D.P. Gunn?Week 4: Latoya Johnson—Can you tell me the title of [author]'s dissertation? He finished his degree at [institutionname] in 1964. Do you have it in your library? How can I obtain a copy of this dissertation?Week 5: Rosa Manuz—How did [sports team name] become the name for [institution name]'s sports teams? Canyou refer me to a book or article that discusses it?Week 6: Chang Su—Repeated question.


Each transaction was classified according to 12 attributes (each with multiple values) and 59categories, initially. Content categories were developed from the data and were also based onthe International Federation of Library Associations (IFLA) digital reference guidelines(International Federation of Library Associations, 2005), the Reference and User ServicesAssociation (RUSA) guidelines for implementing and maintaining virtual reference services(Reference and User Services Association, 2004b), and on the RUSA guidelines for behavioralperformance of reference and information service providers (Reference and User ServicesAssociation, 2004a). After all the transactions were coded by one coder, 10% of the data wascoded by a second coder to evaluate the level of inter-coder reliability. The coding wasfollowed by a discussion among the coders that clarified codes and modified the codingscheme. The final coding scheme was decreased to 23 categories (Table 1). Three iterations ofcoding by the two coders were required until an acceptable level of agreement between the twocoders was reached (above 90%). A different set of transactions (10% of the data) was coded

Table 1Coding scheme

Code Description

First name Librarian addressed the user by first name (e.g. Mary, Moshe)Full name Librarian addressed the user by full name (e.g., Mary Anderson, Moshe

Cohen)With honorific The user's name appeared with honorific (Ms. Mrs.)Greetings Librarians address the user with greetings (Hello, Hi, Greetings)Reiteration of request included inreply

Librarian reiterated user request in reply

Automatic reiteration of requestinclude in reply

Automatic inclusion of user's request in librarian's reply

Article sent Librarian sent the full text of the requested article to the userList of sources suggested Librarian suggested a list of sources on the topic of interestPopulation question answered Librarian provided population numbers for both yearsFull dissertation citation provided Librarian provided the full citation of the dissertationMascot explained Librarian explained the source of the Mascot nameInstructions provided Librarian provided instructions on how to use a resource (ILL, local library,

database)Referral was made Librarian referred the user (Local library, librarian, ILL)Policies explained Librarian explained policies (ILL, opening hours, unaffiliated users, type of

questions answered, copy and download)Search strategy explained Librarian explained search strategy used to find answerEvaluative remarks were given Librarian made evaluative remarks on answerFollow-up Librarian made follow-up remarks (e.g., please let me know if this answers

your question)Name of librarian Librarian signed his/her nameDepartment Librarian signed his/her departmental affiliationContact info Librarian provided contact informationConcluding remarks were made Librarian made concluding remarks (e.g., I hope this is helpful)Thank you Librarian wrote thank you to user at beginning or end of replyApologies Librarian apologized


per iteration. The final result of the inter-coder reliability was 92% (number of agreementsdivided by the sum of number of agreements and disagreement).

Limitations of this study are due mainly to the small sample size, which could not supportcross tabulation among user groups on most content categories. Future studies should increasethe number of institutions and involvemore transactions to determine the statistical significanceof the findings.

5. Results

The content analysis of the 138 e-mail transactions revealed differences in the quality ofservice that virtual reference librarians provide to various users groups. This is the case whenAfrican American and Arab users are compared with Whites, both Christian and Jewish users.Table 2 summarizes the frequencies of codes for each user group. Each of the ethnic orreligious groups is represented by a name; in this section a shorter version of the names will beused; each user group will be indicated by first name of the user (for example, Mary instead ofMary Anderson and Moshe instead of Moshe Cohen).

Fig. 1 illustrates the difference in the average amount of time it took a librarian to respond tousers' requests for each user name. It is very clear that Moshe is getting the quickest reply and

Table 2Code frequencies per user group

Code Mary Moshe Latoya Ahmed Rosa Chang

First name 7 8 9 5 9 5Full name 3 2 2 4 2 6With honorific 5 2 2 0 1 0Greetings 13 15 14 13 16 15Reiteration of request included in reply 2 1 2 1 1 0Automatic reiteration of request include in reply 18 19 22 19 22 20Article sent 0 1 1 2 0 1List of sources suggested 1 0 0 0 0 0Population question answered 4 2 1 3 3 3Full dissertation citation provided 4 3 2 2 2 1Mascot explained 1 2 1 0 5 1Instruction provided 4 3 1 3 1 1Referral was made 10 10 17 10 13 14Policies explained 11 7 6 8 6 4Search strategy explained 1 2 2 3 2 2Evaluative remarks were given 0 0 1 2 0 0Follow-up 4 8 4 4 5 2Name of librarian 14 13 12 13 16 14Department 10 9 6 7 10 11Contact info 5 3 4 5 8 5Concluding remarks were made 6 11 8 6 6 3Thank you 8 6 5 4 7 7Apologies 0 1 3 2 1 2

Fig. 1. Response time (average number of days).


the best level of service. It is also obvious that Ahmed is getting the worst level of service as ittakes the librarian on average much longer to reply to his requests. At this point, it should beclarified that Latoya and Mary can only be compared to one another and not to the other userssince these requests were sent during the weekend. Yet it is clear that Mary's response time isshorter when compared with Latoya's.

Another way to look at these results is by examining the number of days it takes for each ofthe users to get a reply. Fig. 2 describes the number of days it took the librarians to respond to

Fig. 2. Response time (number of days it took the librarian to reply).

Fig. 3. Length of reply.


the user's requests. Again, while Mary and Moshe receive most of the replies within the sameday or during the following day, some of the replies to Latoya and Ahmed are sent weeks afterthey were submitted (some of the messages that Ahmed received arrived on the 31st day aftersending the requests, and Latoya received some of the replies on the 18th and 21st days afterthe requests were sent).

Fig. 3 illustrates the differences among user groups in the length of replies the userreceived from the librarian. The number of words in the reply is an indication of the replylength; for each user the average length of reply was calculated. Again, Moshe and Mary aregetting longer messages than the other user groups and a different level of service is clearlyexhibited.

Fig. 4 illustrates the level of service each user group received in terms of the ratio of thenumber of messages that answered the request to the number of messages the librarian sent theuser. In this figure, the lower the difference between the number of answers and the number ofmessages the higher the level of service the user received. In other words, if the user receivesone response that answers the question, she will be satisfied. On the other hand, when a userreceives many messages but still does not receive a reply to the request, she will not be

Fig. 4. Number of messages received.


satisfied with the service. When comparing users, it is evident that Mary andMoshe receive thehighest level of service in this category while the level of service for Ahmed and Latoya is thelowest. Ahmed and Latoya received many messages but fewer answers, compared with Mosheand Mary, who received proportionally many more answers out of the messages they received.

As Fig. 5 illustrates, overall Mary, Moshe, and Rosa receive more answers to their requests,while Ahmed, Latoya, and Chang do not. This trend is even more clearly represented in Fig. 6,where it can be seen that on almost any type of question Mary gets more answers than doAhmed and Latoya. Overall, most of the time the librarians did not provide the answers toquestions that they were not expected to answer according to their institutional policies andlicensing agreements (e.g., full-text article or topical question). This is described in Fig. 7,which shows many more responses to the first three types of questions (dissertation, mascot,and population) than to the last two types of questions (full-text article and topical).

Another indication of the quality of service is the way the librarian addresses the user. Fig. 8illustrates the different frequencies of greetings, use of first and full name, and honorific use foreach user. Yet again, Ahmed gets the lowest level of service, with no single case of honorificand a lower frequency of first name use.

Along these lines, Fig. 9 describes the frequencies of the professional endings that librariansincluded in their replies to each of the users. As the high bars indicate, Moshe is getting the best

Fig. 5. Answers received by type of question per user.

Fig. 6. Answers received by user per type of question.


level of service. In addition, the ways librarians sign the replies vary across different users.Fig. 10 illustrates this with frequencies bars for each user on each category of signature thatwas included in messages. The two bars in the middle represents lower frequencies of contactinformation, names, and thank you notes for Ahmed and Latoya.

6. Discussion

As the findings show, on all codes and attributes that were examined in this study Latoyaand Ahmed were discriminated against when compared with the level of service Moshe andMary received. Poor service or avoidance of user information needs for long periods of timeare discriminatory behaviors that were evident in the provision of online reference services toArabs and African Americans when compared with White (Caucasian) Jewish or Christianusers.

An examination of these findings in light of RUSA guidelines for behavioral performanceof reference and information service providers (Reference and User Services Association,2004a) and IFLA digital reference guidelines (International Federation of Library

Fig. 7. Answers received by type of question.

Fig. 8. Librarian greetings.


Associations, 2005) is the main focus of the discussion. Quotes from these guidelines illustratethe quality of virtual reference service that is expected from librarians as indicated by theseprofessional associations. On the majority of these service quality aspects, the findings were infavor of Whites (Christian and Jewish users) over African American and Arabs; the quality of

Fig. 9. Professional endings in librarians' messages.

Fig. 10. Closure of librarians' messages.


services to Asians and Hispanics were in between. For example, RUSA guidelines specifythat:

Approachability behaviors, such as the initial verbal and non-verbal responses of the librarian, will set the tone for the entirecommunication process, and will influence the depth and level of interaction between the staff and the patrons. At this stagein the process, the behaviors exhibited by the staff member should serve to welcome the patrons and to place them at ease.The librarian's role in the communications process is to make the patrons feel comfortable in a situation that may beperceived as intimidating, risky, confusing, and overwhelming (Reference and User Services Association, 2004a, Item 1).

The different frequencies of greetings, use of first and full name, and honorific use for eachuser indicate an important aspect of reference service quality. Librarians' approachability,friendliness, and politeness were found by past researchers to affect user's satisfaction(Dewdney & Ross, 1994; Durrance, 1989). On this aspect of service Ahmed gets the lowestlevel of service, with no single case of honorific and lower frequency of first name. The lack ofhonorific could be interpreted as a lack of respect toward Ahmed. IFLA digital referenceguidelines state that librarians should “Show professional courtesy and respect whenanswering questions” (International Federation of Library Associations, 2005). Findingsindicate that librarians respected Moshe and Mary more than they respected Ahmed. At thesame time using first name could indicate friendliness (or lack thereof). Ahmed and Changwere less frequently addressed by first name compared with the other users. It should beclarified, however, that both Chang Su and Ahmed Ibrahim are challenging names.Distinguishing between first name and last name in these two cases is not simple. AhmedIbrahim can as well be Ibrahim Ahmed, and Chang Su can be Su Chang. It is also possible thatin the case of Chang Su a confusion about user gender resulted in librarians addressing the userby full name instead of first name and not using an honorific, which would have to be genderspecific. Even more, Chinese often change the order of first and last name, which makes it


difficult to decide on one of the two. For these reasons, it is possible that the lower amount offirst name use is not entirely an indication of less friendly librarians.

In addition to greeting, IFLA guidelines specify that librarians should structure theirresponses to include a heading, body, and closure as stated in the following quote:

Heading: Greet patron, include a generic notice of thanks for using the service, refer directly to subject of patron'sinquiry.… Body: Cite sources fully, and in a consistent citation style.… Signature: A signature should be a part ofevery closure. The librarian signature may contain librarian name or initials, title, institution and any contactinformation, as is prescribed by supervisor. Examples: “We hope the information we've provided will assist you withyour research”; “I hope you find this information helpful”; “We hope this answers your question. If you have furtherquestions, please contact us again and we will be glad to provide additional assistance…” (International Federation ofLibrary Associations, 2005).

As these guidelines indicate, the ways in which librarians end their replies is an importantaspect of quality service; yet, the endings and signature vary across the different user groups.In the messages to Moshe, librarians follow their professional guidelines for providingeffective reference service and provide more concluding and follow-up remarks (comparedwith the other user groups). In addition, the ways in which librarians sign the replies varyacross different users. It should be stressed that signatures and endings could well be a result ofinstitutional or departmental policies and thus should be applied in a similar manner to allusers. Nonetheless, lower frequencies of contact information, librarians' names, and thank younotes were found in the replies sent to Ahmed and Latoya compared with the other user groups.Thus, Ahmed and Latoya received poor service while Moshe received the highest level ofservice on these aspects.

RUSA guidelines specify that an important aspect of quality reference service is that thelibrarian shows interest in the user's request, even when it is not intellectually stimulating orchallenging. The guidelines specify:

A successful librarian must demonstrate a high degree of interest in the reference transaction.… Librarians whodemonstrate a high level of interest in the inquiries of their patrons will generate a higher level of satisfaction amongusers. To demonstrate interest, the librarian:… Acknowledges user email questions in a timely manner (Reference andUser Services Association, 2004a, Item 2).

Librarians' interest in the users' request is reflected in the response time; it is animportant indication of the quality of the reference service. IFLA guidelines concur andstate that librarians should: “Acknowledge receipt of patron question. Provide patrons withresponses as quickly as possible. Letters and other forms of communication should beanswered promptly and courteously” (International Federation of Library Associations,2005). If users receive a response to their queries quickly, they will be more satisfied withthe service than if the librarian takes a long time to respond. Receiving a reply within theshortest period of time indicates a better level of service. IFLA guidelines continue andstate that librarians should “…create and adhere to stated response turnaround policy”(International Federation of Library Associations, 2005). The expected response time wastwo business days, as stated in most of the studied libraries' response time polices. Thefindings indicate that the average response time to Moshe and Mary is shorter comparedwith Ahmed and Latoya. Similarly, the number of days it takes for each of the users to geta reply indicates a better level of service to Moshe and Mary and a much lower level ofservice to Ahmed and Latoya. It is most likely that a user who receives a reply a month


after asking for information will not be satisfied with the service. The variations in responsetime again indicate discriminatory behaviors of avoidance or poor service. The mostfrequent type of discrimination in public accommodations that was reported by AfricanAmericans involves avoidance or poor service (Feagin, 1991).

Along these lines, it is assumed that if the librarian puts more time into responding to auser request and writes a longer message, a higher quality of response will result. The lengthof the reply (number of words) that Moshe receives is much higher than any other user; Marydoes not get as good service as Moshe, but hers is still better than any of the other users.Moreover, when a user receives many messages but still does not receive a reply to theinformation need, the user will not be satisfied with the service. Again, Moshe and Mary aregetting a higher level of service than the other user groups while the level of service forAhmed and Latoya is the lowest in terms of length of reply and proportion of messages toanswers.

Certain types of questions could potentially trigger different levels of service to varioususer groups while other types of questions will not. It is possible that a librarian will not onlyprefer to help certain user groups more than others but will also avoid declining a requestfrom favorable user groups and at the same time will decline such services from unfavorableuser groups. Bushman and Bonacci (2004), who examined discrimination against Arabs,report that certain individuals are more likely to discriminate against an Arab and to providebad news rather than good news to an Arab. Along these lines, it is possible that a librarianwill behave differently when either providing or refusing to provide a service. It is possiblethat refusing a user's request (e.g., as a response to the request for full-text article) will evokedifferent reactions than helping users. For that reason, this study had two groups of requests,the first three requests (dissertation, mascot, and population), which are expected to beanswered by the librarian, and the next two (full-text article and topical), which are not. Dueto institutional policies and licensing agreements, a librarian was expected to provide ananswer to the first three questions but be disinclined to do so for the next two questions. Theterms of use, in the agreements the library signed with the information owners, usuallyrestrict access to affiliated (and walk in) users. IFLA guidelines further specify that librariansshould “comply with contractual licensing agreements, for both electronic and printmaterials, as well as specific restrictions of use, and any copyright laws governing thematerials in question” (International Federation of Library Associations, 2005). Overall,librarians comply with these guidelines towards all users. It is rare that the librarians providethe full text of an article to unaffiliated users. However, it should be anecdotally mentionedhere that Ahmed received more full-text articles than any other user (a service that he, likethe other users in our study, was not supposed to get as an unaffiliated user). But when hereceived the article in one case (out of two full-text articles that he received), he also receivedanother message from the librarian apologizing for sending him the article that he was notsupposed to get as an unaffiliated user. Despite his receipt of these articles, on almost anytype of question Ahmed gets fewer answers than Mary does (and the same can be said ofLatoya).

Service discrimination against Arabs and African American by reference librarians shouldnot be the norm.While this study has not examined these behaviors at the level of an individual


librarian, the users were discriminated against by individual librarians. Performance evaluationof reference librarians, like other service providers, should also look at how an individuallibrarian treats diversity and, more specifically, assess if a librarian provides unbiased serviceand does not discriminate against users based on race, gender, ethnicity, national origin, orreligion. While this type of assessment in face-to-face interactions might not be feasible, virtualreference services, which archive older transactions, enable such evaluation. In corporateAmerica, such evaluations have been proposed; it was emphasized as a way to facilitatediversity management (Allen, Dawson, Wheatley, & White, 2004; Swanson, 2002). “Themanagers' performance reviews should include how well they manage diversity. This task ofreinforcement and maintenance … falls upon the managers and the respective styles that theymodel for everyone else in the group” (Swanson, 2002, p. 259). They “should be willing to holdsubordinates accountable for achieving these objectives… [a] major issue of control is holding asubordinate accountable for violations of diversity policies” (Allen et al., 2004, p. 14). A librarythat would like to improve virtual reference services equality can hide user names fromlibrarians and insert the names back (automatically) into the response that is sent to the user.This can be further facilitated by using a different field to insert first name and last name to avoidconfusionwhen dealingwith unfamiliar or other challenging names. A library that would like toimprove equality of services can:

• Increase the awareness of librarians to their subjective bias through training. This can bedone through routine diversity training sessions for virtual reference librarians and diversitytraining for librarians in general.

• Managers of reference librarians and researchers could include service equality in theevaluation of virtual reference services.

• Managers can include equality of services as part of a performance evaluation.• LIS schools should emphasize the importance of providing equitable services to differentuser groups online and face-to-face.

• Libraries and LIS schools should make more efforts to recruit minorities to the libraryprofession.

The discriminatory behavior of virtual reference services could be explained by the flaws ofthe virtual environment, where it is easier to behave in less socially accepted ways. It may aswell be explained by other factors, such as the relative homogeneity of librarians. The libraryprofession is overwhelmingly White, homogeneity that does not reflect the diversecomposition of its users (Adkins & Espinal, 2004). This composition does not providemany opportunities for direct contact with colleagues from different backgrounds and providesan ideal atmosphere for stereotypes, prejudices, and discrimination to flourish. One of theways that the libraries can improve the level of service to diverse users groups is by increasingthe efforts to recruit underrepresented groups in the library profession. However, this is not aneasy task because “a significant portion of librarians consider libraries racist workingenvironments” (Curry, 1994, p. 305).

Generalizations about these findings should be made with caution. Although the findingsshow that librarians provide different levels of service to different users on all the aspects that


have been examined, the sample size is not large enough for generalization, and future studiesshould examine more transactions to determine if these differences are statistically significant.Future research should also focus on reference equality at the physical reference desk becauseit is possible that reference librarians provide unequal levels of service only online. Further, itis possible that other online services (in the for profit sectors) will exhibit similar tendenciesbut at the same time it is possible that they will exhibit more or less inequality in their service.Likewise, other library services should be examined to determine equality of services. Forexample, the level of service that librarians provide to affiliated users might be higher andpotentially more equitable. It is also possible that real time services and collaborative virtualreference services will exhibit different levels of equality.

Other research directions that future studies could aim at would be: Are there anydifferences in the geographical location of libraries and equitability of their services? Wouldthe same result be found in public libraries or other academic institutions (not ARL libraries)?Are different attitudes evident between public and private institutions of higher education? Doservices in other countries follow the same equitable pattern (codes of ethics for librariansaround the globe discuss equitable service, but practices may differ)? Can one expect requestsfrom different nationalities to be answered with/without bias? Can one expect that the level ofaccuracy of responses will be affected by this bias as well? These are only some of thequestions that this study raises for future research.

7. Conclusion

Virtual reference librarians discriminated against Arabs and African Americans andprovided the best level of service to Caucasians (Christian and Jewish). This discriminatorypattern was evident on all the dimensions of service quality that were evaluated. Librariansignored requests made by Arabs and African American users more than other users, respondedmore slowly to their requests, put less time and effort into the reply, and when replying to theseusers' requests did not adhere to professional guidelines as much as they did when replied toCaucasian users' requests.

Do librarians exemplify a new type of racism in their virtual reference services? Librariansmay not be actively endorsing racist activities, but they are in control of a great deal of powerthrough their ability to provide complex services and answer user requests over the Internet.Are librarians abusing their power by controlling and censoring access to information based onrace, gender, religion, or national origin? The findings presented above suggest that they may,even if not intentionally.

References

Abels, E. G. (1996). The e-mail reference interview. Reference Quarterly, 35(3), 345−358.Adkins, D., & Espinal, I. (2004). The diversity mandate. Library Journal, 129(7), 52−54.Allen, R., Dawson, G., Wheatley, K., & White, C. S. (2004). Diversity practices: Learning responses for modern

organizations. Development and Learning in Organizations, 18(6), 13−15.


American Library Association (1995). Code of ethics of the American Library Association. Retrieved February 20,2005, from http://www.ala.org/ala/oif/statementspols/codeofethics/codeethics.htm

Bushallow-Wilbur, L., DeVinney, G., &Whitcomb, F. (1996). Electronic mail reference service: A study. ReferenceQuarterly, 35(3), 359−371.

Bushman, B. J., & Bonacci, A. M. (2004). You've got mail: Using e-mail to examine the effect of prejudicedattitudes on discrimination against Arabs. Journal of Experimental Social Psychology, 40(6), 753−759.

Carter, D. S., & Janes, J. (2003). Unobtrusive data analysis of digital reference questions and service at the InternetPublic Library: An exploratory study. Library Trends, 49(2), 251−265.

Civil Rights Act of 1964 § 201, 42 USC, § 2000(a) (1964).Coffman, S. (2003). Going live: Starting and running a virtual reference service. Chicago: American Library

Association.Coffman, S., & Arret, L. (2004). To chat or not to chat: Taking yet another look at virtual reference. The Searcher,

12(8), 49−56.Curry, D. A. (1994). Your worries ain't like mine: African American librarians and the pervasiveness of racism,

prejudice and discrimination in academe. Reference Librarian, 45/46, 299−311.Curry, A., & Copeman, D. (2005). Reference service to international students: A field stimulation research study.

Journal of Academic Librarianship, 31(5), 409−420.Dewdney, P., & Ross, C. S. (1994). Flying a light aircraft: Reference service evaluation from a user's viewpoint.

Reference Quarterly, 34(2), 217−230.Douglas, K. M., & McGarty, C. (2001). Identifiability and self-presentation: Computer-mediated, communication

and intergroup interaction. British Journal of Social Psychology, 40(3), 399−416.Durrance, J. (1989). Reference success: Does the 55 percent rule tell the whole story? Library Journal, 114,

31−36.Feagin, J. (1991). The continuing significance of race: Antiblack discrimination in public places. American

Sociological Review, 56(1), 101−116.Glaser, J., & Kahn, K. B. (2005). Prejudice and discrimination and the Internet. In Y. Amichai-Hamburger (Ed.),

The social psychology of the Internet (pp. 247−274). Oxford: Oxford Univ. Press.Gross, M., & Saxton, M. L. (2002). Integrating the imposed query into the evaluation of reference service: A

dichotomous analysis of user ratings. Library & Information Science Research, 24, 251−263.Hernon, P., & Calvert, P. (2005). E-service quality in libraries: Exploring its features and dimensions. Library &

Information Science Research, 27, 377−404.Hernon, P., & McClure, C. R. (1986). Unobtrusive reference testing: The 55 percent rule. Library Journal, 111(7),

37−41.International Federation of Library Associations (2005). IFLA digital reference guidelines. Retrieved January 5,

2006, from http://www.ifla.org/VII/s36/pubs/drg03.htmKaske, N., & Arnold, J. (2002). An unobtrusive evaluation of online real time library reference services. Retrieved

February 20, 2005, from http://www.lib.umd.edu/groups/digref/kaskearnoldunobtrusive.htmlLiu, M., & Redfern, B. (1997). Information seeking behavior of multicultural students: A case study at San Jose

State University. College and Research Libraries, 58, 348−354.McClure, C., Lankes, R. David, Gross, M., & Choltco-Devlin, B. (2002). Statistics, measures and quality

standards for assessing digital library services: Guidelines and procedures. ERIC Clearinghouse. RetrievedDecember 22, 2005, from http://quartz.syr.edu/rdlankes/tiki-download_file.php?fileId=7

Reference and User Services Association (2004a). Guidelines for behavioral performance of reference andinformation service providers. Retrieved January 5, 2006, from http://www.ala.org/ala/rusa/rusaprotools/referenceguide/guidelinesbehavioral.htm

Reference and User Services Association (2004b). Guidelines for implementing and maintaining virtual referenceservices. Retrieved September 20, 2005, from http://www.ala.org/ala/rusa/rusaprotools/referenceguide/virtrefguidelines.htm

Shachaf, P. (2005). Bridging cultural diversity through email. Journal of Global Information TechnologyManagement, 8(2), 46−60.

http://www.ala.org/ala/oif/statementspols/codeofethics/codeethics.htm

http://www.ifla.org/VII/s36/pubs/drg03.htm

http://www.lib.umd.edu/groups/digref/kaskearnoldunobtrusive.html

http://www.ala.org/ala/rusa/rusaprotools/referenceguide/guidelinesbehavioral.htm

http://www.ala.org/ala/rusa/rusaprotools/referenceguide/guidelinesbehavioral.htm

http://www.ala.org/ala/rusa/rusaprotools/referenceguide/virtrefguidelines.htm

http://www.ala.org/ala/rusa/rusaprotools/referenceguide/virtrefguidelines.htm


Shachaf, P., Hara, N., Meho, L., Kwon, N., Li, B., Sakai, Y., & Lankes, D. (2005). Cross-cultural analysis of virtualreference. [electronic version]. Proceedings of the 68th Annual Meeting of the Society for Information Scienceand Technology, vol. 42.

Sproull, L., & Keisler, S. (1986). Reducing social context cues: Electronic mail in organizational communication.Management Science, 32, 1492−1513.

Stacy-Bates, K. (2003). E-mail reference responses from academic ARL libraries: An unobtrusive study. Referenceand User Services Quarterly, 43(1), 59−70.

Swanson, D. R. (2002). Diversity programs: Attitude and realities in the contemporary corporate environment.Corporate Communications: An International Journal, 7(4), 257−268.

White, M. D. (2001). Digital reference services: Framework for analysis and evaluation. Library & InformationScience Research, 23, 211−231.

Whitlatch, J. B. (1989). Unobtrusive studies and the quality of academic reference services. College and ResearchLibraries, 50(2), 181−194.

Whitlatch, J. B. (2001). Evaluating reference services in the electronic age. Library Trends, 50(2), 207−217.Whitmire, E. (1999). Racial differences in the academic library experiences of undergraduates. Journal of

Academic Librarianship, 25(1), 33−37.Whitmire, E. (2003). Cultural diversity and undergraduates' academic library experiences. Journal of Academic

Librarianship, 29(3), 148−161.Zoe, L. R., & DiMartino, D. (2000). Cultural diversity and end-user searching—An analysis by gender and

language background. Research Strategies, 17(4), 291−305.

are virtual reference services color blind?

Documents